The challenge of offside for VAR
With the introduction of the Video Assistant Referee experiments in 2016, many leagues and competitions, but also technical providers have gained valuable experience in offline and live testing.
One of the major challenges that has emerged is the real-time assessment of offside positions leading FIFA to further investigate this topic.
The provision of a VAR system carried a number of challenges including making all broadcast angles available in real-time to the VAR and the operator in order to make decisions in line with the IFAB’s directive: “maximum benefit, minimum interference”.
Checking offside positions of players has proven to be one of the more difficult tasks due to the number of variables and the very short time-span in which the decision needs to be made. The exact pitch dimensions including any physical camber on the field as well as distortions of the camera lenses make it very difficult for a virtual line to be drawn that accurately represents a true straight line as it would be physically drawn across the field. In addition, with a large number of cameras to choose from, the provider must be ready to show the virtual offside line from the best possible angle in any given situation instantly.
FIFA invited a number of providers to demonstrate their technology and assess the accuracy of such systems in October 2017 using a match day broadcast setup in a Bundesliga 2 stadium in Sandhausen (Germany). The tests showed that a number of VAR system providers are capable of calibrating offside lines from a number of different angles accurately or even very accurately whereby the additional challenge of instant provision will be tested at a later stage.
The experience with this topic has shown how a seemingly straightforward and intuitive issue needs a much more refined solution and that virtual offside lines used in broadcasting should be treated with care if they are being used for potentially match-deciding situations.
Providers were asked to insert offside lines through markers placed at 6 randomly selected locations. We set out a further grid of markers for each case using survey grade equipment and industry leading measurement techniques and asked for a second image. With the cameras in the same position, both images could be superimposed and using the principles of photogrammetry, the parallelity of the offside line to the goal plane could be assessed.
Labosport Ltd working with Sheffield Hallam University designed a series of tests. This included surveying the whole pitch to millimeter accuracy in 3D, and the placement of over 100 calibration points. The calibration points were used as reference markers and the providers had to identify off-side decisions for a range of simulated game situations. The VAR provides had limited time to present the findings between each simulation.